INTRODUCTION
Organizations and work teams abound with emotions, and these emotions can influence key performance outcomes (Ashkanasy, Reference Ashkanasy, Dansereau and Yammario2003; Cole, Walter, & Bruch, Reference Cole, Walter and Bruch2008; Li, Ahlstrom, & Ashkanasy, Reference Li, Ahlstrom and Ashkanasy2010). Contemporary research has defined emotions as complex, multifaceted states that reflect intense affective experiences directed toward a clearly specified target (Frijda, Reference Frijda1993). In work teams, for example, a member may experience specific emotions toward teammate A that differ from the emotions toward teammate B. Research has demonstrated that such interpersonal emotions critically shape members’ interaction and cooperation (Cuddy, Fiske, & Glick, Reference Cuddy, Fiske and Glick2007; Roseman, Reference Roseman, Scherer, Schorr and Johnstone2001). Individuals’ affective attachment to their coworkers therefore represents a key driving force of their performance outcomes (Seers, Petty, & Cashman, Reference Seers, Petty and Cashman1995; Tse, Lam, Lawrence, & Huang, Reference Tse, Lam, Lawrence and Huang2013).
The extant literature provides little knowledge about the dyadic and group-based sources of interpersonal emotions within work teams. Previous research has mainly examined individual characteristics (e.g., personality traits, emotional abilities, and psychological states) as antecedents of interpersonal emotions in work teams (Lourdes & Extremera, Reference Lourdes and Extremera2014), ignoring interpersonal factors. Importantly, however, interpersonal emotions between teammates originate—by definition —within the dyadic relationship between an individual and a specific coworker (Berscheid & Ammazzalorso, 2001; Elfenbein, 2008), and this dyadic relationship, in turn, is embedded within the context of the team as a whole (Lam, Van der Vegt, Walter, & Huang, Reference Lam, Van der Vegt, Walter and Huang2011; Lawler & Thye, Reference Lawler, Thye, Stets and Turner2006). Consequently, dyadic and team-level features should be vital elements in the development of interpersonal emotions.
The few studies addressing this issue have typically used an episodic approach, for example focusing on interpersonal conflicts (Spector, Reference Spector and Cooper1998) or injustice experiences (Weiss, Suckow, & Cropanzano, Reference Weiss, Suckow and Cropanzano1999) that arouse interpersonal emotional reactions. It is not just specific interaction episodes, however, that determine how people respond emotionally toward each other. Scholars have noted that an emotion-eliciting stimulus ‘can also be a stable feature of the work environment’ (Elfenbein, Reference Elfenbein2007: 320). More specifically, the affect theory of social exchange suggests that dyadic relations are a crucial setting in which discrete interpersonal emotions ‘happen to people’ (Lawler & Thye, Reference Lawler, Thye, Stets and Turner2006: 301; see also Lawler, Reference Lawler2001). Dyadic work relations, in particular, involve social exchanges between employees that reflect the quality of the actor-partner relationship (for example, based on a sense of trust and reciprocity within the dyad), with tangible consequences for individuals’ behavioral and affective reactions (Shore, Coyle-Shapiro, Chen, & Tetrick, Reference Shore, Coyle-Shapiro, Chen and Tetrick2009).
Thus, we draw on the affect theory of social exchange to examine the quality of a team member's dyadic relations (i.e., the extent to which an actor perceives the relationship with a specific coworker as mutually trusting, supportive, and based on reciprocity; Blau, Reference Blau1964; Settoon & Mossholder, Reference Settoon and Mossholder2002) that can shape a member's emotions toward a teammate. In particular, we focus on two types of interpersonal emotions: contacting emotions (e.g., admiration and sympathy) and distancing emotions (e.g., envy and contempt), reflecting how individuals are either drawn closer to or pushed away from their coworkers during on-going workplace interactions (Melwani & Barsade, Reference Melwani and Barsade2011; Menon & Thompson, Reference Menon and Thompson2010; Roseman, Reference Roseman, Scherer, Schorr and Johnstone2001).
Consistent with other recent emotion theories (e.g., Elfenbein, Reference Elfenbein2007; Manstead & Fischer, Reference Manstead, Fischer, Scherer, Schorr and Johnstone2001), the affect theory of social exchange further emphasizes that dyadic relationship quality, by itself, is not sufficient to fully explicate an individual's emotional reactions toward others (Lawler, Reference Lawler2001). Relational exchange processes are typically embedded within a larger social context, and characteristics of this context may shape the emotional meaning and significance of a specific interpersonal relation (Lawler & Thye, Reference Lawler, Thye, Stets and Turner2006). In work teams, for example, the way team goals are structured is an important factor that influences members’ interactions and reactions toward each other (Tjosvold, Reference Tjosvold1988). In particular, we focus on the role of cooperative team goals, defined as an emergent team-level construct that reflects members’ shared belief that their goal achievements are interdependent (such that one team member can only reach his or her individual goals when others achieve theirs as well; Deutsch, Reference Deutsch1949; Johnson & Johnson, Reference Johnson and Johnson1989; Wong, Tjosvold, Yu, Reference Wong, Tjosvold and Yu2005). Research has shown that cooperative team goals may set strong norms for cooperation and harmony within the team as a whole (Tjosvold, Yu, & Hui, Reference Tjosvold, Yu and Hui2004) and, thus, may override the salience of relationship quality within a single dyad inside the team (Lam et al., Reference Lam, Van der Vegt, Walter and Huang2011). Accordingly, we cast cooperative team goals as a contextual, cross-level moderator that shapes the role of dyadic relationship quality for interpersonal emotions in team-member dyads.
We examine these ideas across two independent studies conducted in China. Study 1 uses a time-lagged sample of undergraduate students working in project teams, whereas Study 2 employs cross-sectional data from company work teams. Study 2 also investigates the role of interpersonal emotions for employees’ objective work performance, highlighting the practical relevance of understanding the development of such emotions.
Taken together, the present research aims to contribute to the literature on emotions in organizations in several ways. First, we offer new insights into the origins of team members’ interpersonal emotions, focusing on dyadic relationship quality as a key emotion trigger that prior work has largely overlooked. With team-based structures permeating modern organizations and direct, dyadic member relations representing a key feature of teamwork (Kozlowski & Ilgen, Reference Kozlowski and Ilgen2006), such relational aspects are critical to fully understand the development of interpersonal emotions in today's workplace. Second, this study addresses scholars’ repeated calls for a greater contextualization of emotion research (e.g., Elfenbein, Reference Elfenbein2007; Gooty, Gavin, & Ashkanasy, Reference Gooty, Gavin and Ashkanasy2009) by investigating cooperative team goals as a moderator. Integrating dyadic (relationship quality) and structural (cooperative team goals) sources of interpersonal emotions in work teams, we offer a novel, more comprehensive approach to understanding how specific interpersonal emotions emerge within team member dyads. Finally, given that relational concerns (guanxi) are a core element of work and social life in China (Chen, Chen, & Huang, Reference Chen, Chen and Huang2013; Leung, Reference Leung, Huang and Bond2012; Xin & Pearce, Reference Xin and Pearce1996), this research broadens the literature on emotions in Chinese organizations (cf. Li et al., Reference Li, Ahlstrom and Ashkanasy2010). We would expect that dyadic relationship quality and mutual cooperative interdependence play a particularly salient role in shaping interpersonal emotions in China.
THEORETICAL BACKGROUND AND HYPOTHESES
Dyadic Relationship Quality and Interpersonal Emotions
Interpersonal emotions represent feeling states that are directed toward another individual (Frijda, Reference Frijda1993). This type of feelings has been categorized into ‘contacting’ emotions that bring people closer toward each other and entail feelings of attachment, versus ‘distancing’ emotions that drive people away from each other and entail feelings of distinctiveness and contrast (Melwani & Barsade, Reference Melwani and Barsade2011; Menon & Thompson, Reference Menon and Thompson2010; Roseman, Reference Roseman, Scherer, Schorr and Johnstone2001). In this study, we consider two discrete contacting emotions: admiration (i.e., appreciation of the target) and sympathy (i.e., compassion for the target), and two discrete distancing emotions: envy (i.e., having a blind desire for what the target has) and contempt (i.e., devaluing and looking down upon the target; Fischer & Roseman, Reference Fischer and Roseman2007; Lazarus, Reference Lazarus1991). We focus on these specific emotions because previous studies have illustrated them as fundamental reactions toward individuals’ social perceptions (e.g., Cuddy et al., Reference Cuddy, Fiske and Glick2007), indicating the possible relevance of dyadic social relations for these feeling states. In addition, these emotions have been associated with important interpersonal and intergroup behaviors (e.g., helping and harming, Cuddy et al., Reference Cuddy, Fiske and Glick2007), potentially facilitating or disrupting members’ cooperation with their teammates and, thus, shaping their performance outcomes (Alexander, Brewer, & Hermann, Reference Alexander, Brewer and Hermann1999; Weiner, Reference Weiner1986).
The affect theory of social exchange provides a useful framework for understanding the role of dyadic relationship quality for discrete interpersonal emotions (Lawler & Thye, Reference Lawler and Thye1999; Reference Lawler, Thye, Stets and Turner2006). On a general level, social exchange theory argues that individuals tend to reciprocate the positive or negative outcomes they receive in dyadic relations (Blau, Reference Blau1964; Gouldner, Reference Gouldner1960). This can refer to tangible resources (e.g., goods or rewards) but also to immaterial aspects (e.g., goodwill and attitudes; Cropanzano & Mitchell, Reference Cropanzano and Mitchell2005). Specifying this notion, the affect theory of social exchange suggests that dyadic social exchanges can trigger attribution processes that crucially shape an actor's views and evaluations of a dyadic partner (Lawler, Reference Lawler2001; Lawler & Thye, Reference Lawler, Thye, Stets and Turner2006). Eventually, these attributions give rise to discrete interpersonal emotions, such as the ones examined in the present research (Lawler & Thye, Reference Lawler, Thye, Stets and Turner2006; Weiner, Reference Weiner1986).
High-quality relations with a fellow teammate, for example, are based on perceptions of positive, favorable interpersonal exchanges (e.g., a fair, reciprocal distribution of resource investments and outcomes between the interaction partners; Shore et al., Reference Shore, Coyle-Shapiro, Chen and Tetrick2009), and they are therefore likely to lead a focal team member to process information about the target in favorable ways (Baumeister & Leary, Reference Baumeister and Leary1995)—triggering perceptions of the target as friendly, sincere (Cuddy et al., Reference Cuddy, Fiske and Glick2007), deserving of positive outcomes, and undeserving of negative outcomes (Laham, Reference Laham2009; Weiner, Reference Weiner1986). More specifically, research has shown that individuals in a high-quality relationship are likely to attribute an interaction partner's successes to internal aspects, such as the partner's hard work, efforts, and abilities (Feather, Reference Feather1999; Lawler & Thye, Reference Lawler, Thye, Stets and Turner2006). In contrast, individuals are likely to attribute an interaction partner's negative outcomes to external factors beyond the partner's control in high-quality relations, such as bad luck or situational difficulties (Feather, Reference Feather1999; Laham, Reference Laham2009; Lazarus, Reference Lazarus1991). Importantly, these favorable attributions closely map onto the two contacting emotions examined in the present research. As Cuddy et al. (Reference Cuddy, Fiske and Glick2007) noted, admiration is based on the perception that others deserve their positive outcomes, whereas sympathy is based on the perception that a target's negative outcomes are undeserved and beyond the target's control. Consequently, the attributional tendencies described above are likely to guide individuals to more frequently experience interpersonal admiration and sympathy toward the respective target in high-quality relations. Hence, we hypothesize,
Hypothesis 1: In team member dyads, the actor's perceived relationship quality with the target will be positively related to the actor's admiration of the target.
Hypothesis 2: In team member dyads, the actor's perceived relationship quality with the target will be positively related to the actor's sympathy toward the target.
Whereas these hypotheses predict similar associations between dyadic relationship quality and admiration and sympathy as distinct contacting emotions, we expect the role of relationship quality to differ markedly for the two distancing emotions examined in this study (i.e., envy and contempt). In particular, we expect envy to occur more frequently in high- rather than low-quality relations. This prediction may appear counter-intuitive on first glance, as envy is often seen as a negative interpersonal emotion that ‘involves coveting another's superior outcome and comprises feelings of injustice and inferiority’ (Cuddy et al. Reference Cuddy, Fiske and Glick2007: 634). Recent research, however, has depicted envy as more complex and ambiguous (Tai, Narayanan, & Mcallister, 2010). In particular, envying a target's achievements implies that these achievements are favorably recognized and acknowledged (Cuddy et al., Reference Cuddy, Fiske and Glick2007). As such, it is clear that envy also comprises positive perceptions and assessments, which as outlined above, are more likely to occur within high-quality interpersonal relations (Mussweiler & Ruter, Reference Mussweiler and Ruter2003).
Moreover, scholars have emphasized that envy—by its very nature—is an emotion that results from social comparison processes, such that envy ‘follow[s] from workers’ comparisons of desired outcomes relative to those of their coworkers’ (Greenberg, Ashton-James, & Ashkanasy, Reference Greenberg, Ashton-James and Ashkanasy2007: 33; see also Menon & Thompson, Reference Menon and Thompson2010; Vecchio, Reference Vecchio2005). A broad body of research suggests that such social comparisons occur more frequently within high-quality social relations (e.g., Marsh, Reference Marsh1987; Mussweiler & Ruter, Reference Mussweiler and Ruter2003). Specifically, actors have been shown to perceive other individuals as more salient comparison targets to the extent that these others are seen as psychologically proximal (and thus, relatively similar) to the actor (Festinger Reference Festinger1954; Salovey & Rodin, Reference Salovey and Rodin1984; Tesser, Millar, & Morre, Reference Tesser, Millar and Moore1988). Logically, then, individuals strive for social equity with others that are psychologically close, potentially developing a desire for these target's positive outcomes (Pritchard, Reference Pritchard1969; Smith & Kim, Reference Smith and Kim2007). As noted before, high-quality interpersonal relations are based on mutual trust, supportiveness, and reciprocity (Blau, Reference Blau1964; Settoon & Mossholder, Reference Settoon and Mossholder2002) and, consequently, an actor is likely to perceive psychological closeness to a target within such relations (Tesser, Reference Tesser, Millar and Moore1988). While acknowledging the target's achievements and viewing these achievements in a positive light, the actor therefore is likely to frequently compare him- or herself with the target and yearn for similarly favorable outcomes, thus developing envy toward the target.
Within low-quality interpersonal relations, by contrast, an actor is less likely to recognize, acknowledge, and appreciate a target's successes (Laham, Reference Laham2009; Lawler, Reference Lawler2001; 1995; Setton & Mossholder, Reference Settoon and Mossholder2002). Moreover, the actor is likely to perceive the target as rather distal and dissimilar in this situation, with the target therefore appearing less salient for the actor's social comparison processes (Tesser et al., Reference Tesser, Millar and Moore1988). Consequently, within low-quality dyadic relations, we expect that there is less opportunity for interpersonal envy to arise. Hence, we hypothesize:
Hypothesis 3: In team member dyads, the actor's perceived relationship quality with the target will be positively related to the actor's envy toward the target.
In contrast to envy, we expect interpersonal contempt to occur less frequently within high-quality relationships and more frequently within low-quality relationships. Low-quality relations, in particular, are characterized by social exchanges that suffer from a lack of mutual trust, support, and reciprocity (Setton & Mossholder, Reference Settoon and Mossholder2002). Accordingly, such unfavorable relations are likely to negatively color the processing of information about the target (Lawler, Reference Lawler2001). Therefore, in a low-quality relationship, an individual is likely to develop unfavorable perceptions of and attributions towards the target.
More specifically, research suggests that individuals tend to emphasize an interaction partner's failures and shortcomings in low-quality relations (Lawler & Thye, Reference Lawler, Thye, Stets and Turner2006; Weiner, Reference Weiner1986) and to attribute such misfortunes to internal, controllable factors (e.g., a lack of ability, motivation, or effort), thus blaming the target him- or herself for these negative outcomes (Roseman, Reference Roseman, Scherer, Schorr and Johnstone2001; Frijda, Reference Frijda1993). These unfavorable perceptions and attributions closely map onto the conceptual meaning of interpersonal contempt. Such contempt arises when a target's negative attributes and outcomes are perceived as ‘onset controllable’ and, thus, within the target's own responsibility (Cuddy et al., Reference Cuddy, Fiske and Glick2007: 634), such that the target is perceived as deserving of his or her failures and misfortunes (Melwani & Barsade, Reference Melwani and Barsade2011). Consequently, the attributional tendencies within low-quality interpersonal relations are likely to induce contempt, a feeling unlikely to be experienced in the context of a high-quality interpersonal relationship.
Hypothesis 4: In team member dyads, the actor's perceived relationship quality with the target will be negatively related to the actor's contempt toward the target.
The Moderating Role of Cooperative Team Goals
The affect theory of social exchange further emphasizes that the context within which dyadic relations are embedded can decisively influence the development of interpersonal emotions (Lawler & Thye, Reference Lawler and Thye1999; see also Elfenbein, Reference Elfenbein2007). Lawler and Thye (Reference Lawler and Thye1999: 224), for instance, have suggested that ‘emotions are socially constructed . . . in the context of the various social roles, memberships, identities or categories that individuals occupy’. Accordingly, overarching task structures and interdependencies within a team may shape the role of dyadic relations for individuals’ interpersonal perceptions and attributions and the associated emotional reactions (Lawler, Reference Lawler2001). Hence, we consider cooperative team goals and the resulting team-level outcome interdependencies (Deutsch, Reference Deutsch1949; Johnson & Johnson, Reference Johnson and Johnson1989) as a potential contextual boundary condition of the linkage between dyadic relationship quality and interpersonal emotions within work teams.
Research has shown that cooperative team goals generate a focus on common outcomes, motivating members to achieve team goals (Tjosvold et al., Reference Tjosvold, Yu and Hui2004). Given a shared interest, cooperative goals are likely to trigger members’ identification with the team as a whole (Deutsch, Reference Deutsch1949; Johnson & Johnson, Reference Johnson and Johnson1989). Research has demonstrated that such shared group identification can activate broad egalitarian and cooperative norms within the overall team, soften perceived distinctiveness in the team, and diminish differences in individuals’ emotional reactions toward specific teammates (Lawler & Yoon, Reference Lawler and Yoon1998). Thus, the potential relevance of dyadic relationship quality for the development of discrete interpersonal emotions should be less pronounced with highly cooperative team goals. Theorists have argued that ‘the salience of contextual influences of group goals is greater than the salience of interpersonal processes’ (Hogg & Hardie, Reference Hogg and Hardie1992: 42; see also Hogg, Reference Hogg1991). Hence, we expect that cooperative team goals will serve as an important moderating factor, weakening the associations between dyadic relationship quality and interpersonal emotions.
With shared team processes and goals taking precedence over dyadic relations, members will focus their attention on their team's overall outcomes and, thus, will be less attentive towards specific social exchanges in their dyadic relations with other teammates. Therefore, individual team members’ perceived positive or negative outcomes and the associated attributions should become less salient as a basis for discrete interpersonal emotions. Under the influence of cooperative goals, team members are likely to identify with their team as a whole and to view their team in a positive light (Lawler & Yoon, Reference Lawler and Yoon1998). Under this condition, positive dyadic relationship quality may do little to further increase experiences of interpersonal admiration, sympathy, and envy toward individual teammates. Similarly, unfavorable dyadic relationship quality is less likely to promote feelings of interpersonal contempt in this situation. All in all, we propose that cooperative team goals will buffer the positive linkage between dyadic relationship quality and interpersonal admiration, sympathy, and envy as well as the negative linkage between relationship quality and interpersonal contempt.
With less cooperative goals, in contrast, team members should focus on common outcomes to a limited extent because different members’ goal achievement is largely independent (Johnson & Johnson, Reference Johnson and Johnson1989). Also, a strong sense of identification with the overall team is less likely to develop (Tjosvold et al., Reference Tjosvold, Yu and Hui2004). Under this circumstance, team members’ interpersonal emotions should, to a large extent, derive from evaluations of dyadic relationship quality and from the perceptual and attributional tendencies triggered within such relations (Lawler, Reference Lawler2001; Lawler & Thye, Reference Lawler, Thye, Stets and Turner2006).
Hypothesis 5: In team member dyads, the actor's perceived relationship quality with the target will be more strongly positively related to the actor's admiration (5a), sympathy (5b), and envy (5c) toward the target in teams with less cooperative goals than in teams with more cooperative goals.
Hypothesis 6: In team member dyads, the actor's perceived relationship quality with the target will be more strongly negatively related to the actor's contempt toward the target in teams with less cooperative goals than in teams with more cooperative goals.
Interpersonal Emotions and Individual Task Performance
Theorists have noted that emotions serve adaptive functions to guide individuals’ behavior (e.g., Frijda, Reference Frijda1993; George, Reference George2011). More specifically, a team member's contacting and distancing emotions at work may provide salient signals for the potential value of social interactions with other members (Lawler & Thye, Reference Lawler, Thye, Stets and Turner2006), shaping the members’ behavior toward his or her teammates and, thus, influencing the social resources the individual can utilize for task accomplishment. We expect that a team member's task performance will benefit if he or she, on average, experiences more contacting emotions (admiration and sympathy) toward fellow teammates, whereas task performance should decline for members that, on average, experience more distancing interpersonal emotions (envy and contempt).
As noted before, a member who frequently experiences admiration toward his or her teammates acknowledges others’ successes and positive outcomes and attributes these benefits towards teammates’ competencies, knowledge, and efforts (Cuddy et al., Reference Cuddy, Fiske and Glick2007). As such, the member is likely to closely associate with fellow teammates to benefit from their superior qualities (Sweetman, Spears, Livingstone, & Manstead, Reference Sweetman, Spears, Livingstone and Manstead2013). Further, when frequently experiencing interpersonal sympathy, a member recognizes other teammates’ failures and negative outcomes while, at the same time, acknowledging that these problems are undeserved and beyond targets’ control (Feather & Sherman, Reference Feather and Sherman2002). Again, the member is likely to closely associate with fellow teammates in this situation, in an effort to help others to overcome their misfortunes and improve their outcomes (Cuddy et al., Reference Cuddy, Fiske and Glick2007). Consistent with this reasoning, research has shown that contacting emotions facilitate approach behavior (Cacioppo & Gardner, Reference Cacioppo, Gardner and Spence1999), prompting individuals to actively cooperate with others in a friendly and productive manner (Alexander et al., Reference Alexander, Brewer and Hermann1999; Cuddy et al., Reference Cuddy, Fiske and Glick2007).
Hence, although admiration and sympathy constitute discrete contacting emotions, we argue they are both likely to increase a focal member's social resources within his or her work team. By maintaining and extending positive interactions with fellow teammates (Côté, Reference Côté2005), a member should be able to draw upon others’ knowledge, advice, and tangible help if needed. This may create unique performance potentials by building an enduring pool of social resources that can promote a member's task accomplishment (Fredrickson, Reference Fredrickson2001).
Hypothesis 7: A team member's average level of admiration (7a) and sympathy (7b) toward teammates will be positively related to his or her individual task performance.
The role of distancing emotions, namely envy and contempt, for individual task performance may be more ambiguous. On the one hand, theorists have suggested that negative emotions may provide distinct benefits in some situations, for example, aiding individuals’ creative performance by strengthening their task persistence (de Dreu, Baas, & Nijstad, Reference de Dreu, Baas and Nijstad2008). Envy, in particular, may enhance individuals’ work motivation so as to reduce perceived self-other performance discrepancies (Schaubroeck & Lam, Reference Schaubroeck and Lam2004). Also, contempt may trigger perceptions of strength and superiority, thus motivating individuals’ sustained work efforts to affirm and maintain their social status (Keltner & Haidt, 1999).
On the other hand, research has linked negative affective experiences with detrimental performance outcomes (e.g., Kaplan et al., Reference Kaplan, Bradley, Luchman and Haynes2009; Staw & Barsade, Reference Staw and Barsade1993). We believe these dysfunctional consequences will be particularly pronounced for distancing emotions within a team context. After all, both envy and contempt may provide salient signals that deter close interactions and trigger a tendency to avoid other teammates (Cacioppo & Gardner, Reference Cacioppo, Gardner and Spence1999). A member frequently experiencing envy is likely to feel inferior to his or her teammates, triggering defensive behaviors and withdrawal tendencies to avoid further threats to one's self-image (Cuddy et al., Reference Cuddy, Fiske and Glick2007; Vecchio, Reference Vecchio2005). Similarly, a member experiencing contempt toward teammates emphasizes other's failures and blames teammates themselves for these unfavorable outcomes (Frijda, Reference Frijda1993). The member is likely, then, to distance him- or herself from other team members and to exclude them from social interactions (Melwani & Barsade, Reference Melwani and Barsade2011). Consequently, both envy and contempt are likely to isolate a member within his or her team (Melwani & Barsade, Reference Melwani and Barsade2011; Vecchio, Reference Vecchio2005). Research has shown that distancing interpersonal emotions can hamper cooperation and create substantial conflict among team members (Cuddy et al., Reference Cuddy, Fiske and Glick2007; Lazarus, Reference Lazarus1991).
Consistent with this argumentation, we suggest that frequent experiences of interpersonal envy and contempt will diminish the pool of social resources available to a focal team member, as the respective member is less likely to benefit from others’ knowledge, information, and tangible help when performing his or her tasks. Rather, the focal member may devote substantial resources towards avoiding and/or dealing with unpleasant interactions and conflicts with other teammates (Lazarus, Reference Lazarus1991; Pugh, Reference Pugh, Lord, Klimoski and Kanfer2002). As such, he or she is likely to be distracted from core job tasks, potentially deteriorating his or her performance outcomes (Frijda, Reference Frijda1986).
Hypothesis 8: A team members’ average level of envy (8a) and contempt (8b) toward teammates will be negatively related to his or her individual task performance.
METHOD
We tested the hypotheses in two independent studies conducted in China. In Study 1, we collected temporally lagged data from student project teams. In Study 2, we collected cross-sectional data from work teams in a telecommunication services company. We note that some of the survey data were also used in previous research (Lam et al., Reference Lam, Van der Vegt, Walter and Huang2011; Tse et al., Reference Tse, Lam, Lawrence and Huang2013).
Study 1 – Sample and Procedures
We collected survey data from a sample of undergraduate (second- and third-year) students in an education institute in Macau, China, to test Hypotheses 1 to 6. As a part of their course requirements, participants worked for 3 months in teams of 3 to 7 members to complete a team project that required intense member interaction (viz., preparing a business plan). After 1.5 months (Time 1), we collected data on relationship quality and the control variables. One month later (Time 2), we measured cooperative team goals and interpersonal emotions. We distributed paper-and-pencil surveys to 154 students in 32 teams. Participants returned completed surveys directly to the researchers, and confidentiality was assured.
To measure relationship quality and interpersonal emotions, we collected data at the dyadic level of analysis, using a round-robin design in which each member provided ratings on all other members of the team (Warner, Kenny, & Stoto, Reference Warner, Kenny and Stoto1979). Cooperative team goals were measured at the individual level and then aggregated to the team level of analysis using a referent-shift consensus composition model (Chan, Reference Chan1998; Tjosvold et al., Reference Tjosvold, Yu and Hui2004). The final data set comprised 141 students (555 dyadic relations) within 30 teams (92% response rate). Respondents’ mean age was 20 years; 77% were female; and average dyadic tenure (i.e., the time two members forming a dyad had known each other) was 19 months.
Study 2 – Sample and Procedures
Study 2 was designed to constructively replicate Study 1 and to extend the study by examining the role of interpersonal emotions for individual task performance (Hypotheses 7 and 8). We collected cross-sectional survey data from teams of 4 to 5 sales associates in one of the biggest telecommunication services companies in China. Individual participants’ job duties included both sales tasks and office support. To fulfill these duties, each team member had to navigate multiple interpersonal relations with teammates and to closely coordinate job tasks and responsibilities.
We distributed paper-and-pencil surveys to 132 sales associates within 31 teams. Participants directly returned the completed surveys to the researchers. Confidentiality was assured. Dyadic relationship quality, interpersonal emotions, and cooperative team goals were measured with the same instruments as in Study 1. Further, we drew on company records to obtain objective performance data for each participant. The final data set contained 128 individuals (408 dyadic relations) across 31 teams (97% effective response rate). Respondents’ mean age was 25 years (SD = 5.2 years), and their mean organizational tenure was 1.6 years; 44% of the respondents were female, and 44% had high-school level education or above. Average dyadic tenure (i.e., the time two members forming a dyad had worked together) was 5 months.
Measures
All survey measures were translated to Chinese using a double-blind back-translation procedure. Following previous research using round robin designs that require respondents to rate items with regard to multiple other individuals (e.g., De Jong, Van der Vegt, & Mollenman, Reference De Jong, Van der Vegt and Molleman2007; Venkataramani & Dalal, Reference Venkataramani and Dalal2007), we used shortened measures for our study variables to ameliorate survey fatigue. All of these measures were validated in a pilot study in an organization operating cosmetic chain stores in Hong Kong, China. In the pilot study, we randomly paired respondents within work teams and asked them to complete measures of relationship quality (Graen & Uhl-Bien, Reference Graen and Uhl-Bien1995), cooperative team goals (Tjosvold et al., Reference Tjosvold, Yu and Hui2004), interpersonal emotions (Eisenberg et al., Reference Eisenberg, Fabes, Miller, Fultz, Shell and Mathy1989; Fiske, Cuddy, Glick, & Xu, Reference Fiske, Cuddy, Glick and Xu2002; Parrott & Smith, Reference Parrott and Smith1993), and positive and negative affectivity (used as control variables; Watson, Clark, & Tellegen, Reference Watson, Clark and Tellegen1988). A total of 238 usable questionnaires were returned (89% response rate); 95% of the respondents were female, and 95% had a high school education or above. Mean age and organizational tenure were 25.5 and 2.7 years, respectively. We used two criteria to select items for the current study. First, we conducted a confirmatory factor analysis (CFA) to examine the factor structure of the measures, and we only retained items if they had relatively high (i.e., > 0.70) and statistically significant loadings on their intended factor. Second, we omitted one cooperative team goals item (‘our team members want each other to succeed’) based on face validity concerns in the present main studies’ context. This item seemed unsuitable for the student teams in Study 1 because students’ overall class grade was based on a forced ranking system, such that students may not have wanted each other to be particularly successful within the class as a whole. The resulting, shortened measures are described in more detail below. In the pilot study, Cronbach's alphas for all of these shortened measures were above 0.75. The correlation coefficients between the shortened measures used in the main study and the full version measures were all greater than 0.92 (p < 0.001).
Relationship quality
Following previous research that has studied dyadic relations between team members (e.g., Anderson & Williams, Reference Anderson and Williams1996; Sherony & Green, Reference Sherony and Green2002), we adapted items from Graen and Uhl-Bien's (Reference Graen and Uhl-Bien1995) leader-member exchange measure, such that respondents independently assessed their relationship quality with each of their teammates. Based on the pilot study, five items were chosen for this measure (sample item: ‘How would you characterize your working relationship with this team member?’; 1 = extremely ineffective; 5 = extremely effective). Cronbach's alpha was 0.91 in both Study 1 and Study 2.
Interpersonal emotions
We used emotional adjectives to capture discrete interpersonal emotions (e.g., Feather & Sherman, Reference Feather and Sherman2002; Weiss et al., Reference Weiss, Suckow and Cropanzano1999; 1 = never; 5 = always). Based on the pilot study, we measured admiration with 3 items from Fiske et al. (Reference Fiske, Cuddy, Glick and Xu2002; sample item: ‘admiring’; α = 0.91/0.86 in Studies 1 and 2, respectively), sympathy with 3 items from Eisenberg et al. (Reference Eisenberg, Fabes, Miller, Fultz, Shell and Mathy1989; sample item: ‘sympathetic’; α = 0.90/0.88), envy with 3 items from Parrott and Smith (Reference Parrott and Smith1993) and Fiske et al. (Reference Fiske, Cuddy, Glick and Xu2002; sample item: ‘jealous’ α = 0.82/0.70), and contempt with 3 items from Fiske et al. (Reference Fiske, Cuddy, Glick and Xu2002; sample item: ‘disgusted’; α = 0.91/0.87). Participants independently assessed their feelings with regard to each of their teammates on these items.
Cooperative team goals
Cooperative team goals were measured using three items from Tjosvold and colleagues’ (Reference Tjosvold, Yu and Hui2004) scale. The items were ‘Our team members’ goals go together’; ‘Our team members “swim or sink” together’; and ‘Our team members seek compatible goals’ (1 = strongly disagree; 5 = strongly agree). Cronbach's alpha was 0.88 in Study 1 and 0.87 in Study 2. As indicated before, cooperative goals were conceptualized as a team-level variable, so we aggregated individuals’ responses to the team level. There was significant between-team variance in members’ ratings of cooperative team goals (Study 1: F29, 136 = 9.11, p < 0.001; Study 2: F30, 127 = 7.58, p < 0.001) and interrater reliability and agreement surpassed common standards (ICC1 = 0.62/0.62, ICC2 = 0.89/0.87, median rwg(j) = 0.95/0.87), indicating that aggregation to the team level was justified (Hofmann, Reference Hofmann and Rogelberg2008).
Individual task performance
We obtained individual monthly sales records (number of telecommunication services packages sold) for all respondents in Study 2 during the month in which we conducted the survey. Comparable information on individual members’ task performance was not available in Study 1 since the students received a team grade for the project.
Control variables
We controlled for actors’ and targets’ gender and age, dyadic tenure, and team size (Richter, West, van Dick, & Dawson, Reference Richter, West, van Dick and Dawson2006). Moreover, we included actors’ positive and negative affectivity (i.e., the stable tendency to experience positive vs. negative emotions) as covariates to avoid biasing effects from individuals’ affective dispositions. Based on the pilot study, we used four items from Watson et al. (Reference Watson, Clark and Tellegen1988) to measure both positive (sample item: ‘enthusiastic’; α = 0.86/0.78 in Studies 1 and 2) and negative affectivity (sample item: ‘nervous’; α = 0.75/0.70).
Statistical Analyses
The round-robin design used in this research required that each member of a team rate and be rated by every other member. Therefore, the present data have a complex, nested structure, with individuals nested both within dyadic relationships and within teams. Hence, we employed Kenny and colleagues’ social relations model (using the MLwiN computer package; Goldstein et al., Reference Goldstein, Rasbash, Plewis, Draper, Browne and Yang1998) to test Hypotheses 1 to 6 (Kenny, Reference Kenny1994; Snijders & Kenny, Reference Snijders and Kenny1999). This analytical approach regards each individual both as an actor and as a target and provides random estimates that indicate how much of the variance in a dependent variable (i.e., interpersonal emotions) is explained by characteristics of the actor, the target, the actor-target dyad, and the team. Furthermore, it provides fixed estimates comparable to unstandardized regression coefficients. We first calculated “null models” for each of the four discrete interpersonal emotions that did not contain any predictor variables and were used as a reference for subsequent analyses. These null models also provided an overview of how the variance in interpersonal emotions was partitioned across levels of analysis. Next, we added the control variables, main effects, and two-way interaction terms to examine our hypotheses. We tested for a decrease in log-likelihood between each of these models by means of a chi-square difference test, evaluating the significance of improvements in model fit. Finally, we used the SPSS mixed model procedure to examine Hypotheses 7 and 8 at the individual level of analysis while allowing for random intercepts to control for possible team-level effects (Bickel, Reference Bickel2007). All predictor variables were standardized prior to the analyses (Aiken & West, Reference Aiken and West1991).
RESULTS
Descriptive Statistics and Confirmatory Factor Analyses
Table 1 presents descriptive statistics and bivariate correlations for all measures across both Studies 1 and 2. We first conducted confirmatory factor analyses to evaluate the discriminant validity of the four interpersonal emotion measures. Results suggested that the hypothesized four-factor model (Study 1: CFI = 0.98, TLI = 0.97, RMSEA = 0.07, χ2 = 163.04, df = 48; Study 2: CFI = 0.96, TLI = 0.95, RMSEA = 0.07, χ2 = 138.18, df = 48) yielded a good fit to the data and provided better fit than a one-factor model (Study 1: CFI = 0.40, TLI = 0.26, RMSEA = 0.30, ∆χ2 = 2646.08, ∆df = 6, p < 0.001; Study 2: CFI = 0.35, TLI = 0.21, RMSEA = 0.27, ∆χ2 = 1459.28, ∆df = 6, p < 0.001).
Table 1. Means, standard deviations, and intercorrelations of measures
Note. Correlations for Study 1 appear above the diagonal (N = 555 dyads); correlations for Study 2 appear below the diagonal (N = 408 dyads). For correlations including work performance, N = 128 individuals.
* p < .05. **p < .01. *** p < .001.
Additional confirmatory factor analyses examined the discriminant validity of all six focal variables in our model (i.e., relationship quality, four interpersonal emotions, and cooperative goals). Results showed that the hypothesized six-factor model (Study 1: CFI = 0.97, TLI = 0.96, RMSEA = 0.06, χ2 = 311.79, df = 109; Study 2: CFI = 0.95, TLI = 0.93, RMSEA = 0.07, χ2 = 294.71, df = 109) yielded a good fit to the data and provided substantially better fit than an alternative one-factor model (Study 1: CFI = 0.39, TLI = 0.32, RMSEA = 0.23, ∆χ2 = 4579.85, ∆df = 15, p < 0.001; Study 2: CFI = 0.42, TLI = 0.35, RMSEA = 0.20, ∆χ2 = 2610.14, ∆df = 15, p < 0.001). Overall, these findings supported the discriminant validity of our study variables.
Variance Partitioning
Table 2 shows the partitioning of variance in interpersonal emotions between the actor, target, dyad, and team levels. These findings indicate that, across both of our studies, a substantial portion (ranging from 35% to 61%) of the variance in interpersonal emotions directed toward another teammate depended on characteristics of the dyadic relation between actor and target. Hence, our conceptual focus on the dyadic level of analysis appears substantively meaningful.
Table 2. Variance partitioning for actors’ interpersonal emotions toward targets
Note. Study 1: N = 141 individuals in 555 dyads within 30 teams. Study 2: N = 128 individuals in 408 dyads within 31 teams.
Tests of Hypotheses
Table 3 (Model 1) presents the results for Hypotheses 1 to 4, which predicted specific associations between perceptions of dyadic relationship quality and interpersonal emotions. Across both studies, such relationship quality was positively associated with actors’ admiration (Study 1: B = 0.36, p < 0.001; Study 2: B = 0.29, p < 0.001), sympathy (Study 1: B = 0.21, p < 0.001; Study 2: B = 0.21, p < 0.001), and envy (Study 1: B = 0.14, p < 0.001; Study 2: B = 0.13, p < 0.001) and negatively associated with actors’ contempt toward the target (B = −0.07, p < 0.05; Study 2: B = −0.13, p < 0.001), even after accounting for the control variables. Hypotheses 1, 2, 3, and 4 were therefore supported.
Table 3. Quality of social relationships and actors’ interpersonal emotions toward targets
Note. Study 1: N = 555 dyads. Study 2: N = 408 dyads. *p < .05. **p < .01. *** p < .001.
Table 3 (Model 2) presents the results for Hypotheses 5 and 6, which predicted that associations between relationship quality and interpersonal emotions are stronger in teams with less rather than more cooperative goals. As shown, the cross-level interaction coefficients of relationship quality and cooperative team goals were significant for all interpersonal emotions in both Study 1 (admiration: B = −0.08, p < 0.01; sympathy: B = −0.09, p < 0.01; envy: B = −0.07, p < 0.05; contempt: B = 0.09, p < 0.01) and Study 2 (admiration: B = −0.07, p < 0.05; sympathy: B = −0.09, p < 0.01; envy: B = −0.11, p < 0.001; contempt: B = 0.06, p < 0.05). Figure 1 shows these interactions and the results of the simple-slopes analyses for Study 1 (Aiken & West, Reference Aiken and West1991).Footnote [1] As shown, relationship quality was more strongly positively associated with interpersonal admiration, sympathy, and envy in teams with less rather than more cooperative goals, supporting Hypotheses 5a, 5b, and 5c. Further, dyadic relationship quality was more strongly negatively associated with contempt under conditions of less rather than more cooperative goals, supporting Hypothesis 6.
Figure 1. Two-way interaction of relationship quality and cooperative team goals with interpersonal emotions (Study 1)
We tested Hypotheses 7 and 8 in Study 2, which predicted an individual team member's average levels of contacting and distancing emotions to relate with his or her performance. First, we aggregated the participants’ dyadic ratings of interpersonal emotions to the individual level. There was significant between-person variance in interpersonal emotions, including admiration (F127, 407 = 6.46, p < 0.001), sympathy (F127, 407 = 5.39, p < 0.001), envy (F127, 407 = 3.22, p < 0.001), and contempt (F127, 407 = 2.57, p < 0.001). Further, both intraclass correlation coefficients (admiration: ICC1 = 0.52, ICC2 = 0.70; sympathy: ICC1 = 0.47, ICC2 = 0.51; envy: ICC1 = 0.24, ICC2 = 0.76; contempt: ICC1 = 0.31, ICC2 = 0.70) and within-individual agreement indices (admiration: median rwg(j) = 0.94; sympathy: median rwg(j) = 0.97; envy: median rwg(j) = 0.96; contempt: median rwg(j) = 0.98; cf. Bliese, Reference Bliese, Klein and Kozlowski2000) supported aggregation to the individual level of analysis.
We regressed individuals’ objective sales records on all four of these individual-level emotions at the same time, controlling for team size and respondents’ age, gender (0 = male, 1 = female), and organizational tenure. As shown in Table 4, both envy (B = −0.23, p < 0.01) and contempt (B = −0.12, p < 0.05) were negatively related to sales performance, whereas there were no significant relationships for admiration (B = −0.02, ns) and sympathy (B = 0.07, ns). Hence, Hypothesis 7 was rejected whereas Hypothesis 8 was supported.
Table 4. Multilevel analysis for individual task performance (Study 2)
Note. † p < .10, *p < .05, **p < .01, *** p < .001. N = 128 individuals.
DISCUSSION
Using data from two independent samples, the present studies demonstrate the role of dyadic relationship quality for the development of four discrete interpersonal emotions between team members (i.e., admiration, sympathy, envy, and contempt). Furthermore, we uncovered cooperative team goals as an important contextual moderator, such that the role of relationship quality for interpersonal emotions was more pronounced in teams with less cooperative goals and buffered in teams with more cooperative goals. Finally, at the individual level of analysis, Study 2 found negative relationships between distancing interpersonal emotions (viz., envy and contempt) and objective task performance.
Theoretical Implications
These results contribute to a better understanding of the origins of interpersonal emotions in the workplace by focusing on individuals’ evaluations of dyadic relationship quality as a key explanatory variable. Previous studies on this issue have tended to focus (explicitly or implicitly) on appraisals of particular interaction episodes or incidents (e.g., Spector, Reference Spector and Cooper1998; Weiss & Cropanzano, Reference Weiss, Cropanzano, Staw and Cummings1996). By contrast, this research illustrates that interpersonal emotions may develop in the context of one's ongoing interpersonal relationships within a work team (cf. Elfenbein, Reference Elfenbein2007; Lawler & Thye, Reference Lawler, Thye, Stets and Turner2006). As such, beyond specific incidents or events giving rise to interpersonal emotions, such feelings also may emerge in the course of team members’ day-to-day social interactions.
Second, this study contributes to the emotion literature by providing evidence that experiences of interpersonal emotions in dyadic relationships depend on contextual factors (Lawler & Thye, Reference Lawler and Thye1999; Manstead & Fischer, Reference Manstead, Fischer, Scherer, Schorr and Johnstone2001). Cooperative team goals, in particular, can influence team members’ emotions toward others, to the point of overriding the consequences of dyadic relationship quality assessments. These findings show that the combination of relational and structural elements within teams is an important origin of interpersonal emotions. Hence, the current study responds to calls for a greater contextualization of organizational behavior research in general and research on emotions in organizations in particular (Gooty et al., Reference Gooty, Gavin and Ashkanasy2009; Johns, Reference Johns2006). It is interesting to note, in this respect, that our studies have shown cooperative team goals to neutralize the potential role of dyadic relationship quality assessments for a diverse set of both contacting and distancing interpersonal emotions. As such, it appears that interpersonal emotionality is generally less pronounced in teams with more cooperative goals. One may speculate that such contexts emphasize collective team processes and outcomes over dyadic aspects, as the teams’ overall performance is critical for individual members’ goal attainment (Tjosvold et al., Reference Tjosvold, Yu and Hui2004). Hence, emotionality in such teams may be more strongly focused on the team as a whole (e.g., team affective tone; Cole et al., Reference Cole, Walter and Bruch2008) rather than dyadic member relations. Clearly, this notion extends beyond the present studies’ focus, and future research is needed to examine its viability.
Third, the present study focused on discrete interpersonal emotions rather than global positive or negative affect. We found similar associations between relationship quality and the two positive, contacting emotions (admiration and sympathy), but different associations for the two negative, distancing emotions (envy and contempt). This pattern of results is in line with the notion that negative emotions are more diverse (e.g., envy contains both positive and negative sentiments whereas contempt is clearly hostile; Tai et al., 2010; Melwani & Barsade, Reference Melwani and Barsade2011) than positive ones and may even act “in opposition to one another” (Mayer, Gaschke, Braverman, & Evans, Reference Mayer, Gaschke, Braverman and Evans1992: 120). Also, it illustrates the value of focusing on discrete rather than global interpersonal emotions to gain a finer-grained understanding of emotional experiences in organizations (Gooty et al., Reference Gooty, Gavin and Ashkanasy2009).
Fourth, drawing on objective performance outcomes in a real-world setting, the present research corroborates previous work on the performance consequences of emotions in teams and organizations (e.g., Cole et al., Reference Cole, Walter and Bruch2008; Kaplan et al., Reference Kaplan, Bradley, Luchman and Haynes2009). Notably, our findings of relevant performance linkages for envy and contempt, but not for admiration and sympathy, support the notion that negative emotions may have more powerful consequences than positive ones (Baumeister, Bratslavsky, Finkenauer, & Vohs, Reference Baumeister, Bratslavsky, Finkenauer and Vohs2001). Scholars have linked envy and contempt, for example, to various psychological and physiological outcomes, such as depression, withdrawal cognitions, and anger as well as reduced learning intentions and deteriorated health (e.g., Fischer & Roseman, Reference Fischer and Roseman2007; Melwani & Barsade, Reference Melwani and Barsade2011; Vecchio, Reference Vecchio2005). These adverse effects may distract individuals from core task pursuit and debilitate an individual's performance potentials (Menon & Thompson, Reference Menon and Thompson2010; Tai et al., Reference Tai, Narayanan and McAllister2012). The roles of admiration and sympathy, in contrast, may be more indirect, as we have theorized these contacting emotions to strengthen individual performance by triggering positive social behaviors that invite others’ cooperation and helping (Sweetman et al., Reference Sweetman, Spears, Livingstone and Manstead2013; Weiner, Reference Weiner1986). Given our non-significant findings for admiration and sympathy, we encourage future research to further explore the performance consequences of these discrete contacting emotions by examining potential indirect, mediated relationships. Further, such work may also consider alternative performance dimensions (e.g., organizational citizenship, counterproductive behaviors, and creative performance), rather than exclusively focusing on core task performance (cf. Borman & Motowidlo, Reference Borman and Motowidlo1997; de Dreu et al., Reference de Dreu, Baas and Nijstad2008; George, Reference George2011).
Finally, our findings also contribute to emotion research in China. Most of the existing studies on this issue have investigated the impacts of emotions on employees’ well-being, job attitudes, and behavior (Li et al., Reference Li, Ahlstrom and Ashkanasy2010) and have examined how specific cultural aspects (e.g., strong collectivism) may shape employees’ emotional experiences and expressions (Suh, Diener, Oishi, & Triandis, Reference Suh, Diener, Oishi and Triandis1998). Importantly, scholars have distinguished two elements of collectivism, including a concern for relationships with specific others, on the one hand, and a concern for relationships with the group as a whole, on the other (Brewer & Chen, Reference Brewer and Chen2007). The present findings illustrate the interplay of such specific (i.e., dyadic relations) and group-based (i.e., cooperative team goals) factors as emotional antecedents, with relational features losing their relevance for interpersonal emotions when group-based features become salient. We encourage future research to further examine the role of cultural aspects for interpersonal emotion emergence in teams, focusing on the potential dominance of group-based features in the context of Chinese collectivistic values, in particular.
Practical Implications
From a practical perspective, the results of our two studies suggest that managers could take active steps to facilitate contacting emotions (i.e., interpersonal admiration and sympathy) by encouraging high-quality relationships between team members. To do that, managers may consider deep-level variables in team composition, such as team members’ agreeableness, extraversion, and emotional stability, which have been shown to facilitate high quality relationships between team members in prior studies (Bell, Reference Bell2007).
Even in well-composed teams, however, it may not be possible to maintain high-quality relationships at all times, given that teamwork is often rife with potentials for conflict and disagreement (Levine & Moreland, Reference Levine and Moreland1990). Further, as shown in our research, even high-quality relationships may generate distancing emotions such as envy that may damage task performance. Our research findings suggest the usefulness of cooperative team goals to address this issue. Managers may achieve this by setting collective targets for the team as a whole, giving feedback and rewards based on the team's overall performance, and articulating a common vision that emphasizes team members’ collective identity (cf. DeShon, Kozlowski, Schmidt, Milner, & Wiechmann, Reference DeShon, Kozlowski, Schmidt, Milner and Wiechmann2004). Stressing such cooperative goals may allow managers to prevent distancing emotions among their team members, even when members have less-than-optimal relationship quality.
We believe these practical implications are particularly important for Chinese management. The Chinese society is highly relational-oriented, such that individuals tend to strongly appreciate and value interpersonal relationships (Leung, Reference Leung, Huang and Bond2012). This study shows that a singular focus on high-quality relations could be misguided even in a Chinese context, since high-quality relations may diminish contempt but may enhance envy between team members. The use of cooperative goals appears critical to prevent both kinds of negative emotions arising from interpersonal relations.
Limitations and Future Research Directions
The present study has some potential limitations associated with its research design. First, following much of the existing research on coworker dyads (e.g., De Jong et al., Reference De Jong, Van der Vegt and Molleman2007; Venkataramani & Dalal, Reference Venkataramani and Dalal2007), we captured our key variables through shortened measures to minimize survey fatigue. We acknowledge that this may raise concerns about measurement validity. At the same time, we note that (a) all items were adapted from well-established scales, (b) we based our choice of items on an independent pilot study, and (c) confirmatory factor analyses in both of our focal studies supported the convergent and discriminant validity of the shortened measures. Moreover, the constructive replication of key results across two samples provides greater confidence in our findings and mitigates possible measurement concerns. Nevertheless, further replication using the full measures may be beneficial in future research.
Second, the present use of self-report measures may introduce the possibility of common-method bias (Podsakoff, MacKenzie, Lee, & Podsakoff, Reference Podsakoff, MacKenzie, Lee and Podsakoff2012). However, several strengths of our research design (e.g., temporally lagged design in Study 1; use of aggregated data for cooperative team goals; constructive replication of key findings; objective measurement of sales performance) alleviate such concerns. And third, the present findings preclude causal conclusions, given our studies’ correlational nature. Future work based on experimental or longitudinal data is required to fully address this concern.
Beyond addressing limitations, the present study suggests several interesting directions for future inquiry. Scholars might build on our findings to develop a more complete model of the emergence of interpersonal emotions in work teams. For example, we did not capture some of the mechanisms theorized to underlie the associations between relationship quality and interpersonal emotions (e.g., actor's attributions, perceived deservingness of the target's outcomes). Incorporating these mechanisms could contribute to a more nuanced understanding of the respective linkages. Similarly, future work could include additional moderating variables that may shape the role of relationship quality. For example, actors’ trait anger (Douglas & Martinko, Reference Douglas and Martinko2001) or trait hostility (Judge, Scott, & Ilies, Reference Judge, Scott and Ilies2006) may negatively color their general perceptions of others, thus overriding the potential influence of specific relationship quality appraisals.
Further, whereas we have concentrated on the moderating role of teams’ cooperative goals, some studies have cast cooperation and competition as distinct variables with differential effects. For example, Tjosvold, Johnson, Johnson, and Sun (Reference Tjosvold, Johnson, Johnson and Sun2003) suggested that competition may exist even within a generally cooperative context. Therefore, future research may benefit from simultaneously examining how both cooperative and competitive team goals affect members’ interpersonal emotions.
Moreover, future research could examine the role of relationship quality for emotions other than the ones examined in the present investigation. Such work could, for example, concentrate on self-focused emotions (e.g., pride, shame) rather than interpersonal emotions (e.g., admiration, contempt). This might offer new insights into the multilevel nature of emotions within work teams (Ashkanasy, Reference Ashkanasy, Dansereau and Yammario2003). As illustrated in this study, dyadic relationships play an important role for interpersonal (i.e., other-focused) emotions. Self-focused emotions, however, may reflect individuals’ affective reactions towards their overall pattern of social relations within the team (Kang, Shaver, Sue, Min, & Jing, Reference Kang, Shaver, Sue, Min and Jing2003). Hence, whereas a dyadic perspective appears justified to examine interpersonal emotions, an individual-level perspective may be required to understand the role of coworker relations for self-focused emotions. This might point towards a theoretically intriguing lack of multilevel homology (Chen, Bliese, & Mathieu, Reference Chen, Bliese and Mathieu2005) in the emergence of different types of emotions.
CONCLUSION
In sum, this study demonstrates important antecedents, contingency factors, and consequences of interpersonal emotions in work teams, outlining both why and under which circumstances specific, discrete interpersonal emotions are likely to occur in team-member dyads and illustrating how such feelings relate with employees’ task performance. We hope our findings will stimulate further research on this issue, helping organizations to design effective interventions that facilitate team members’ favorable interactions, promote contacting and reduce distancing emotions between teammates, and, eventually, improve team members’ performance outcomes.
